Integrating Geographic Information Systems (GIS) and Artificial Intelligence (AI) for Predicting Surface Runoff and Environmental Sustainability – A Future Study (2025–2050) of the Haqlan Valley Watershed

Abstract

This study aims to develop a predictive model for surface runoff in the Haqlan Valley basin, western Iraq, by integrating Geographic Information Systems (GIS) and Artificial Intelligence (AI) techniques, using the Random Forest algorithm and the hydrological SCS-CN model. A spatial database was created, covering land cover, soil types, and rainfall intensity for the period 2000–2024. Rainfall events were classified into medium and high intensities, and surface runoff depth and volume were calculated based on Curve Number (CN) values, which reflect soil and vegetation characteristics.

The Random Forest model was trained on fifty cases representing the relationship between rainfall intensity, CN, and runoff volume, then used to predict runoff for the period 2025–2050 under the SSP245 climate change scenario. The results showed high prediction accuracy and identified critical years, such as 2030 and 2045, which are expected to experience the highest rainfall intensities and runoff volumes.

The study identified hydrological risk zones and classified them into three categories (low, medium, high), with proposed environmental sustainability strategies for each, including water harvesting, rainfed agriculture, and vegetation protection. This research provides a scientific tool to support decision-making in water resource management and to mitigate the impacts of climate change in arid and semi-arid regions.

Keywords: Haqlan